As cannolis get more complex, it can be tricky to figure out what exactly the LLM is seeing on a given AI node. [[Logging arrows]] make this easy. Orange, unlabeled arrows are logging arrows. When an AI node has a logging arrow pointing to a content node, all of the messages, as well as the LLM config of that AI node will be written to the content node, formatted nicely. ![[Pasted image 20240718120128.png]] [[System messages]] ← | → [[That's (pretty much) it!]]